Sexual Compulsivity and Gender Stereotypes in Men Who Have Sex with Men: A Cross-Sectional Study in Spain

Abstract

Men who have sex with men are at increased risk of acquiring sexually transmitted infections. Although behavioral and contextual interventions have improved, infections are still spreading. A new focus is needed regarding the influence of sexual compulsivity and gender self-identification. The objective was to explore the relationship between SC and gender stereotype self-identification in MSM in Spain. A cross-sectional study based on an anonymous online survey of a sample of 881 MSM in Spain was performed. A validated Sexual Compulsivity Scale was applied as well as the Bem Sex Role Inventory to identify gender stereotype self-identification. A total of 87.5% of participants identified as gay, while 106 (12%) were sexually compulsive. The Bem Sex Role Inventory showed excellent reliability (0.92 Cronbach’s alpha). Gender stereotype classification showed the undifferentiated category as predominant (n = 325; 36.9%). The androgynous category had higher scores on the Sexual Compulsivity Scale, while the feminine category had lower scores (p < 0.007). Those identifying as androgynous scored higher overall on the Sexual Compulsivity Scale (15; p < 0.001; [12–20]) and subscales (“Interference of sexual behavior” (5; p < 0.001; [4–7]) and “Failure to control sexual impulses” (9; p = 0.014; [7–12])). Education level, cohabitation, sexual role, and unprotected sexual practices are related to sexual compulsivity according to different gender stereotypes. Sexual compulsivity is related to gender stereotype self-identification in men who have sex with men. Specific sociodemographic, behavior, and sexual profiles exist for those at greater risk of sexually transmitted infection dissemination. The undifferentiated category is at greater risk of engaging in unprotected practices. New avenues of prevention that include these findings could shift heteronormative standards to better understand current trends in the sexual health of men who have sex with men.

1. Introduction

Sexually transmitted infections (STIs) remain a significant public health concern, especially among men who have sex with men (MSM) [1,2]. In addition to the spread of human immunodeficiency virus (HIV) transmission [3], in recent years, other epidemic outbreaks of hepatitis A [4,5] and monkeypox [6,7] have significantly affected the MSM population and have posed a challenge for public health surveillance and prevention efforts in developed countries [8].

Sexual compulsivity (SC) has been found to be significantly associated with factors that determine sexual risk behaviors [9], driving negative health conditions [10]. Sexual compulsivity is defined by sexual preoccupation and a lack of sexual impulse control [11,12] and is characterized by an increase—in intensity, frequency, and duration—of inappropriate or excessive sexual fantasies, such as excessive pornography, the excessive use of the internet for sexual purposes, excessive masturbation, or multiple casual sex partners [13]. Several studies have identified a strong link between SC and risky sexual behavior among MSM, such unprotected intercourse [14], substance use [15,16], and worse mental health [17].

MSM comprise a diverse group whose sexual health and sexual behaviors are often influenced by gender identity, specifically masculinity norms [18]. These norms refer to the set of cultural expectations and behaviors that are typically associated with being male, female, or gay in society; these can vary across cultures and historical periods and should be taken into account in research to address their unique health care needs [19,20]. Previous research reported a relationship between SC and masculine identity [21,22,23]. In the arena of sexual health and STI prevention in MSM, masculine identity has been considered an important aspect, although to date, several gaps have been found due to the lack of consideration of affective and sexual diversity and their relation to this identity, which reduces the effectiveness of preventive interventions [24].

In MSM, the relationships among risk taking and masculinity are complex and paradoxical, since the diversity inherent in this group shapes different strategies to respond to gender demands, such as an alignment with feminine stereotypes, traditional masculine norms, or renewed gay norms [25]. It has been found that traditional masculine norms are related to risk behaviors in MSM, such drug consumption [26] and barebacking [27], and also in specific populations, such as in Black and Latino MSM [28,29,30], but little is known about the impact of the dynamic and non-hegemonic gender identities in MSM [31].

Focusing on the manifested diversity and on the need to develop effective strategies close to real MSM experiences, it is necessary to know how the different categories of gender can suppose more vulnerable situations [32]. Gender is a cultural construct that, when taken as a social structuration system, influences subjectivity and organizes social interactions, as well as influencing sexual encounters, behaviors and ways of relating [31].

Sex and gender are influenced by each other—they are “intertwined”—and some efforts to measure them have been found [33]. The Bem Sex Role Inventory (BSRI) [34] conceptualizes gender self-identification beyond exclusive feminine and masculine traditional stereotypes [35], since it includes categorization based on instrumental (stereotypes socially perceived as masculine) and expressive traits (stereotypes socially perceived as masculine (instrumental traits) or feminine (stereotypes socially perceived as feminine)), providing a four-based categorization: masculine, feminine, undifferentiated, and androgynous. Previous studies have found an association between better health outcomes and androgynous categories [36,37] and worse outcomes in undifferentiated categories, including among MSM [38]. Despite the limitations of the BSRI [39], to date, no measurement tools have been found to analyze the gender stereotypes in MSM beyond traditional masculinity [40].

According to current research, MSM might be at increased risk for HIV and other STIs due to behavioral and contextual factors [41,42] but also because of the influence of sexual compulsivity [14,43] and gender self-identification [38]. Given the limited research on gender stereotypes and SC in the MSM population, our objective was to explore the relationship between SC and gender stereotype self-identification in MSM in Spain.

2. Materials and Methods

A cross-sectional study, based on an online survey of gender stereotypes and sexual compulsivity in a sample of MSM population in Spain, was conducted. A convenience sampling method was used. The subjects were men or individuals who self-identified as men and who met three criteria: they must (1) be between 16 and 75 years old, (2) be residents of Spain, and (3) declare that they have had sexual relations with other men in the past 12 months.

An anonymous online survey was developed and used to collect the data. An ad hoc online questionnaire was designed and disseminated in 2020. Participants were recruited through social networks, such as Facebook, with the support of non-governmental organizations dedicated to HIV prevention in Spain. In addition, participants were also requested from Grindr, a mobile geo-social application, mainly aimed at a gay audience, which enables its users to locate and communicate with gay and bisexual men and women, as well as transgender individuals and transsexuals, in the vicinity. Through their program “Grindr for equality” (G4E), they disseminated the advertisements for the research study for 48 h, encouraging users to participate in the survey. The survey stated that participants could only respond if they were of legal age, which, for sexual matters in Spain, is 16 years old. Anonymity and confidentiality of participants were guaranteed during the process.

2.1. Sociodemographic Variables, Behavioral Characteristics, and Unsafe Sexual Practices

Basic information on the participants, including age, place of birth (Spain, other), education level (secondary school and below, vocational training, or university studies), employment (employed, unemployed, student, or retired), monthly income (≥€ 1200, € 700–1200 and ≤€ 700) and cohabitation situation (stable partner, alone, or other), was collected.

Questions regarding behavioral characteristics and unsafe sexual practices were asked. Information regarding sexual orientation (heterosexual, bisexual, homosexual, or other), sexual role (active/insertive, versatile, or passive/receptive), sexual partners in the past 12 months (regular, causal, or both), use of dating apps (never or sometimes/usually), chemsex participation (never or sometimes/usually; defined as intentional sexual practices in a group while under the effect of psychoactive drugs for a prolonged duration), sex under the influence of drugs (no or yes), PrEP use for HIV (never or sometimes/usually), and previous diagnosis of STIs (no or yes) was also collected. Other sexual practices without protection were included and dichotomized into “never” and “sometimes/usually”: black-kissing (sucking of the anus), face-sitting (sitting on the face), felching (sucking semen ejaculated into the anus), introduction of the fingers or objects into the anus, and fisting (introduction of the fist or part of forearm into the anus).

2.2. Sexual Compulsivity Scale

The translated Spanish version of the Sexual Compulsivity Scale (SCS) of Kalichman et al. (1994) [44] was administered to assess sexual compulsivity. The SCS was previously validated in the MSM population in Spain [45]. This scale consists of 10 Likert-type items, measured on a scale ranging from 1 (not at all) to 4 (always). Participants rate each item ranging from 1 point (not at all like me) to 4 points (very much like me). The total score ranges between 10 and 40 points and is obtained by adding the ratings assigned to each item. Higher scores indicate higher levels of sexual compulsivity. The factorial analysis of the Spanish version resulted in two factors: (1) Interference of sexual behavior, which includes questions about the interference that sex has on several life aspects of the participants (0.968 internal consistency); and (2) Failure to control sexual impulses, which includes items that refer to concerns and difficulties that people have or experience about issues related to sex and its impulses (0.977 internal consistency).

2.3. Gender Stereotype Self-Identification: The Bem Sex Role Inventory

In order to measure gender stereotype self-identification, a reduced version of the popular Bem Sex Role Inventory [46], translated into Spanish [47], was administered. The BSRI-12 measures the gender roles using 12 Likert-type items, measured on a scale ranging from 1 (strongly disagree) to 7 (totally agree). Using these items, participants rated how much different personality characteristics were “like me”. From these, separate scores were calculated for instrumental (masculine) and expressive (feminine) traits and for undifferentiated and androgynous traits.

3. Statistical Analysis

A univariate descriptive analysis was previously performed to assess the frequency and distribution of the variables included in the study. Frequency and percentages were presented for categorical variables whereas the mean and standard deviation (SD) as well as the median and interquartile range (IQR) were used to describe quantitative variables.

To determine SC, participants were dichotomized by the median value. Bivariate analyses were performed to identify significant associations between sexual compulsivity and the behavioral characteristics of the participants. The Kolmogorov–Smirnov test (sample > 50) was applied to check the normality of the quantitative variables, and medians with 25th–75th percentiles were calculated for variables with a non-normal distribution. Nonparametric tests were used to analyze the relationship between the total scores of the SCS and the characteristics of the sample of MSM. The Mann–Whitney U test for two different groups, the Kruskal–Wallis test for more than two groups, and Spearman´s correlation coefficient for continuous variables were used in the bivariate analysis, accepting a statistical significance level of p < 0.05.

To assess the BSRI-12 construct validity in MSM, an exploratory factor analysis of principal components and Varimax rotation with Kaiser correction was performed [48]. Reliability and the internal consistency were analyzed using Cronbach’s alpha test [49]. The Bartlett sphericity test was significant (p < 0.05), and the Kaiser–Meyer–Olkin (KMO) measure of sampling adequacy was higher than 0.75 [50]. The principal factors that resulted in eigenvalues higher than 1 were retained. To assign each item to factors, their factor loadings equal to or greater than 0.5 were considered. Then, the median-split method was used to classify the gender roles [51]. The variable was dichotomized from the median of each of the two factors to obtain the four categories of the variable gender roles proposed: feminine (score above the median for factor 1 and below the median for factor 2); masculine (score above the median for factor 2 and below for factor 1); androgynous (score above the median for both factors); and undifferentiated (score below the median for both factors) [34]. Cronbach’s alpha was used to assess the internal reliability.

To explore the SC relationship with gender stereotype self-identification, a bivariate inferential analysis was performed using contingency tables by means of Pearson’s chi-squared association test, with a statistical significance level of p < 0.05. Then, to assess SC according to gender role self-identification, the Kruskal–Wallis test was used to analyze the median (MED) values, and interquartile ranges (IQRs) were calculated.

Finally, to determine the sociodemographic and sexual behavior association with SC according to gender stereotype self-identification, several binary logistic regression analyses were performed. Those variables that showed statistical significance in the bivariate analyses were used in the models. Among the estimators obtained, the odds ratio (OR) with its 95% confidence interval (95%CI) was taken into account. To assess the goodness of fit of the final model, X2 significance on the omnibus test was determined, and the Cox and Snell R2 and Nagelkerke R2 were also taken into account. Furthermore, the correctly classified overall percentage was retrieved.

Questionnaires were checked for errors before the data was entered, coded, and analyzed using the Statistical Package for the Social Sciences software ((IBM Corp. Released 2023. IBM SPSS Statistics for Windows, version 24.0. Armonk, NY, USA: IBM Corp.) under the University of Zaragoza license.

Ethical Considerations

The study was reviewed and approved by the Aragón Research Ethics Committee, Spain (CEICA; project identification code PI18/327). The research also fulfilled the rules and ethical principles of the Declaration of Helsinki. Participation in the study was anonymous and voluntary. All participants consented via an online informed consent document before participating in the study. Data were anonymized.

4. Results

4.1. Participants

The sample consisted of 881 MSM aged 16 to 75 years old (M = 33.13; SD = 9.99), and the median age was 32 years [IQR: 25–40]. The composition of sociodemographic characteristics of the group can be seen in

Table 1. Sociodemographic characteristics of the 881 Spanish MSM.

	n	%
Age (years), median [IQR]	32 [25–40]
<35	458	52
35–44	209	23.7
≥45	108	12.3
Place of birth
Other	240	27
Spain	641	73
Education level
Secondary school and below	56	6.4
Vocational training	186	21.1
College degree and above	639	72.5
Employment
Employed	682	77.4
Unemployed	47	5.3
Student	145	16.5
Retired	7	0.8
Monthly income
≥€ 1200	481	54.6
€ 700–1200	219	24.9
≤€ 700	181	20.5
Cohabitation Situation
Stable partner	196	22.2
Alone	308	35
Others (family, shared apartment]	377	42.8

Note: [IQR] = interquartile range.

. The majority of participants were of Spanish origin (641; 73%), had a college degree or above (639; 72.5%) employed (682; 77.4%), earning more than 1200 € per month (481; 54.6%), and lived with family or shared an apartment (377; 42.8%).

4.2. Behavioral Characteristics, Unsafe Sexual Practices, and SCS

Our population study comprised 85.7% participants who identified as gay and 12.5% who identified as bisexual. Sexual role was versatile in 49.1% of the participants, and 40.5% had had casual sexual partners in the past 12 months. Almost all of the participants used dating apps (93.9%), while only 10% declared chemsex participation, and nearly 26% of the participants practiced sex with drugs. The majority of the participants never used PrEP (92.3%), and almost half of them had previous diagnoses of any STI (49%). Regarding unprotected sexual practices, our participants declared black-kissing (68.9%), face-sitting (22.9%), felching (6.6%), rimming or the introduction of fingers (79%), the introduction of other objects into the anus (38%), and fisting (14%). A total of 106 (12%) were sexually compulsive, scoring above 23 points in the SCS scale; in the subscales, 248 (28.1%) revealed interference, and 254 (28.8%) presented failure of control (above the 75% quartile).

The associations between behavioral characteristics, unsafe sexual practices, and the total and subscales of SCS scores are presented in

Table 2. Results of bivariate analysis between the Sexual Compulsivity Scale and subscale scores based on participants’ reported behavioral characteristics and sexual risk practices.

			Sexual Compulsivity Scale		Interference of Sexual Behavior Domain		Failure to Control Sexual Impulses Domain
			Score		Score		Score
	n	%	MED [IQR]	p	MED [IQR]	p	MED [IQR]	p
Sexual Orientation
Heterosexual	9	1	18 [15–35]	0.046	7 [5–15]	0.023	13 [8–19.5]	0.123
Bisexual	110	12.5	15 [12–21]		5 [4–8]		9 [7–13]
Gay/homosexual	755	85.7	14 [11–18]		5 [4–7]		9 [7–12]
Others	7	0.8	18 [11–21]		6 [5–8]		11 [6–13]
Sexual Role
Active/insertive	224	27.7	13 [11–18]	0.056	5 [4–6]	0.057	8 [7–11]	0.100
Versatile	433	49.1	15 [12–20]		5 [4–7]		9 [7–12]
Passive/receptive	204	23.2	14 [12–18.75]		5 [4–7]		9 [7–12]
Sexual Partners (past 12 months)
Regular sexual partners	73	8.3	14 [10–17.50]	0.183	5 [4–5.50]	0.062	8 [7–11]	0.474
Casual sexual partners	357	40.5	14 [12–18]		5 [4–7]		9 [7–12]
Both regular/casual	316	35.9	15 [11–20]		5 [4–7]		9 [7–13]
Dating apps
Never	54	6.1	13 [11–16]	0.046	4 [4–6]	0.007	8 [6.75–11.25]	0.175
Sometimes/usually	827	93.9	14 [11–19]		5 [4–7]		9 [7–16]
Chemsex participation
Never	793	90	14 [11–18]	<0.001	5 [4–7]	<0.001	9 [7–12]	0.004
Sometimes/usually	88	10	17 [13–23.75]		7 [4.25–8]		10.5 [7–14]
Sex while under the influence of drugs
No	654	74.2	14 [11–18]	<0.001	5 [4–6]	<0.001	9 [7–12]	0.031
Yes	227	25.8	15 [12–21]		6 [4–8]		9 [7–13]
PrEP use
Never	813	92.3	14 [11–19]	0.202	5 [4–7]	0.156	9 [7–12]	0.503
Sometimes/usually	68	7.7	15.50 [12–20]		5 [4–8]		9 [7–12]
Previous diagnosis of STIs
No	449	51	14 [11–19]	0.865	5 [4–6]	0.074	9 [7–12]	0.251
Yes	432	49	14 [11–19]		5 [4–7]		9 [7–12]
Black-kissing
Never	274	31.1	14 [11–18.25]	0.148	5 [4–6]	0.002	9 [7–12]	0.697
Sometimes/usually	607	68.9	14 [12–19]		5 [4–7]		9 [7–12]
Face-sitting
Never	679	77.1	14 [11–18]	<0.001	5 [4–7]	<0.001	9 [7–12]	<0.001
Sometimes/usually	202	22.9	16 [12–21]		6 [4–8]		10 [7–14]
Felching
Never	823	93.4	14 [11–19]	0.105	5 [4–7]	0.297	9 [7–12]	0.054
Sometimes/usually	58	6.6	15 [12–20.25]		5 [4–7]		10 [7.75–13]
Introduction of fingers
Never	185	21	14 [11–19]	0.546	5 [4–7]	0.102	9 [7–12]	0.870
Sometimes/usually	696	79	14 [12–19]		5 [4–7]		9 [7–12]
Introduction of objects
Never	546	62	14 [11–18]	0.040	5 [4–7]	0.195	9 [7–12]	0.034
Sometimes/usually	335	38	14 [12–20]		5 [4–7]		9 [7–12]
Fisting
Never	758	86	14 [11–18]	<0.001	5 [4–7]	<0.001	9 [7–12]	0.007
Sometimes/usually	123	14	16 [13–22]		6 [4–8]		10 [7–14]

Note: MED: median; IQR: interquartile range; Chemsex: intentional sexual practices under the effect of psychoactive drugs, performed in a group for a prolonged duration; PrEP: human immunodeficiency virus pre-exposure prophylaxis; STIs: sexually transmitted infections. p: p-value; statistical significance when <0.05.

. In the case of participants who identified as heterosexual and others, a higher score in the SCS (total) and in the “Interference of sexual behavior” domain was observed (p < 0.05). A higher SCS total score and “Interference of sexual behavior” domain score was observed in MSM who used information and communication technologies to find a sexual partner (p < 0.05). MSM who have never participated in chemsex and who have not had sexual activity while under the influence of drugs reported a lower score in the total SCS, “Interference of sexual behavior” domain (p < 0.01), and “Failure to control sexual impulses” domain (p < 0.05). The participants who performed face-sitting and fisting presented a higher score in the SCS total scale, in the “Interference of sexual behavior” domain (p < 0.01), and in the “Failure to control sexual impulses” domain (p < 0.05). Those who performed black-kissing scored higher in the “Interference of sexual behavior” domain (p < 0.05), and those who performed the introduction of objects into the anus scored higher in the total scale and the “Failure to control sexual impulses” domain (p < 0.05).

4.3. Psychometric Properties of BSRI-12

To determine the empirical structure of the instrument, an exploratory factor analysis (EFA) was performed. The EFA of the BSRI-12 instrument showed excellent reliability in the Spanish MSM population, with a Cronbach’s alpha of 0.92 and good application conditions according to the sampling adequacy (KMO of 0.90, and p < 0.01 in Bartlett’s sphericity test). Two main factors appeared that had eigenvalues greater than 1. The screen plot of the eigenvalues shows that, for the EFA, the 2-factor structure best fits the data (Figure 1).

The items included in each factor are shown in

Table 3. Gender role self-identification characteristics (BSRI-12) in the Spanish MSM population.

	Factor 1	Factor 2
Items BSRI-12	Feminine: Expressive Characteristics	Masculine: Instrumental Characteristics
BSRI_1: Gentle	0.850
BSRI_2: Compressive	0.872
BSRI_3: Has leadership abilities		0.833
BSRI_4: Acts as a leader		0.878
BSRI_5: Dominant		0.787
BSRI_6: Tender	0.849
BSRI_7: Warm	0.867
BSRI_8: Affectionate	0.883
BSRI_9: Strong personality		0.656
BSRI_10: Defender of own beliefs	0.678
BSRI_11: Sensitive to others’ needs	0.814
BSRI_12: Makes decisions easily		0.615
Cronbach’s Alpha	0.947	0.857
% Explained variance	57.56	15.07

. The first factor (Feminine: Expressive characteristics) was made up of seven items (1: gentle; 2: compressive; 6: tender; 7: warm; 8: affectionate; 10: defender of own beliefs; and 11: sensitive to others’ needs), obtained scores between 0.678 and 0.883, and explained 57.56% of the total variance, with a Cronbach’s alpha of 0.947. The second factor (Masculine: Instrumental characteristics) included five items (3: has leadership abilities; 4: acts as a leader; 5: dominant; 9: strong personality; and 12: makes decisions easily). Items scored between 0.615 and 0.878, accounting for 15.07% of the total variance, and the Cronbach’s alpha coefficient was 0.857. Masculine traits have a low saturation of “feminine” factors and, conversely, feminine traits score low on factor 2 “masculine” factors.

4.4. Gender Stereotypes and Sexual Compulsivity

After classifying the participants according to the self-perception of gender stereotypes, the undifferentiated category was predominant (n = 325; 36.9%), followed by the androgynous (n = 283; 32.1%), masculine (n = 153; 17.4%), and feminine (n = 120; 13.6%) categories. A statistically significant association between SCS and gender role stereotypes was found. Androgynous participants had higher scores above the median (180; 36.6%), while participants endorsing the feminine category had lower scores on the sexual compulsivity scale (57; 14.7%), with p < 0.001 (

Table 4. Classification of participants according to gender role categories and SCS scores.

Gender Role Categories			Sexual Compulsivity Scale Scores
			Median < 14		Median ≥ 14
	n	%	n	%	n	%	p
Feminine	120	13.6	57	14.7	63	12.8	0.007
Masculine	153	17.4	66	17	87	17.7
Androgynous	283	32.1	103	26.5	180	36.6
Undifferentiated	325	36.9	163	41.9	162	32.9
Overall participants	881	100	389	44.2	492	55.8

Note: p-value: Pearson’s chi-squared association test; significance at p < 0.05.

).

Between-group comparisons of gender stereotypes revealed that participants who self-identified as androgynous scored higher than undifferentiated participants on both the overall SCS (15; [12–20] vs. 13; [11–17]; p < 0.001) and the subscales (“Interference of sexual behavior” (5; [4–7] vs. 5; [4–6]; p < 0.001) and “Failure to control sexual impulses” (9; [7–12] vs. 8; [7–11]; p = 0.014) (

Table 5. Median scores on the Sexual Compulsivity Scale and its domain scores according to gender stereotypes.

					Overall Participants (n = 881)		Androgynous (n = 283; 32.1%)		Feminine (n = 120; 13.6%)		Masculine (n = 153; 17.4%)		Undifferentiated (n = 325; 36.9%)
			Items	Range	MED	[IQR]	MED	[IQR]	MED	[IQR]	MED	[IQR]	MED	[IQR]	p
Sexual Compulsivity Scale			10	10–40	14	[11–19]	15	[12–20]	14	[11–19]	15	[12–20]	13	[11–17]	<0.001
Interference of sexual behavior			4	4–16	5	[4–7]	5	[4–7]	5	[4–7]	5	[4–7]	5	[4–6]	<0.001
Failure to control sexual impulses			6	6–24	9	[7–12]	9	[7–12]	9	[7–12.7]	10	[7–13]	8	[7–11]	0.015

Note: MED: Median; [IQR]: Interquartile range; p-value: Kruskal–Wallis test of statistical significance; significance at p < 0.05.

).

Social characteristics and sexual practices are related to SC and its different dimensions, according to the gender stereotypes with which the participants self-identify, as can be seen in

Table 6. Gender stereotypes, population characteristics, unsafe sexual practices, and sexual compulsivity in MSM.

Gender Stereotypes and Sexual Compulsivity Dimensions	Wald	df	Sig.	OR	95% C.I.
					Lower	Higher
FEMININE COMPULSIVITY (GLOBAL)
UNPROTECTED SEXUAL PRACTICES
Face-sitting	4.286	1	0.038	5.749	1.098	30.109
FEMININE INTERFERENCE
COHABITATION SITUATION
Stable partner	4.348	2	0.114	Reference
Alone	3.639	1	0.056	0.250	0.060	1.039
Others (family, shared apartment]	3.211	1	0.073	0.239	0.050	1.144
SEXUAL ROLE
Active/insertive	5.834	2	0.054	Reference
Versartile	4.485	1	0.034	5.261	1.132	24.459
Passive/receptive	5.339	1	0.021	8.123	1.374	48.011
MASCULINE COMPULSIVITY (GLOBAL)
EDUCATION LEVEL
Secondary school and below	5.727	2	0.057	Reference
Vocational training	0.038	1	0.845	0.866	0.204	3.673
College degree and above	5.315	1	0.021	9.014	1.390	58.434
MASCULINE INTERFERENCE
EDUCATION LEVEL
Secondary school and below	4.053	2	0.132	Reference
Vocational training	3.985	1	0.046	0.313	0.100	0.979
College degree and above	0.279	1	0.597	0.609	0.097	3.836
SEXUAL ROLE
Active/insertive	3.746	2	0.154	Reference
Versatile	2.056	1	0.152	2.365	0.729	7.668
Passive/receptive	3.705	1	0.054	3.652	0.977	13.657
SEXUAL ORIENTATION
Heterosexual	10.287	3	0.016	Reference
Bisexual	0.044	1	0.835	1.429	0.050	40.855
Gay/homosexual	1.305	1	0.253	0.147	0.006	40.855
Others	0	1	1.000	0	0	-
UNPROTECTED SEXUAL PRACTICES
Fisting	3.856	1	0.050	3.435	1.002	11.771
ANDROGYNOUS COMPULSIVITY (GLOBAL)
COHABITATION SITUATION
Stable partner	8.960	2	0.011	Reference
Alone	4.773	1	0.029	0.310	0.108	0.888
Others (family, shared apartment]	8.326	1	0.004	0.215	0.076	0.611
SEXUAL ORIENTATION
Heterosexual	5.750	3	0.124	Reference
Bisexual	3.170	1	0.075	0.003	0.008	1.258
Gay/homosexual	5.404	1	0.020	0.061	0.006	0.645
Others	0	1	0.999	0	0	-
ANDROGYNOUS INTERFERENCE
UNPROTECTED SEXUAL PRACTICES
Chemsex	3.912	1	0.048	2.708	1.009	7.266
UNDIFFERENTIATED COMPULSIVITY (GLOBAL)
UNPROTECTED SEXUAL PRACTICES
Face-sitting	7.442	1	0.006	3.611	1.435	9.083
UNDIFFERENTIATED INTERFERENCE
UNPROTECTED SEXUAL PRACTICES
Sex under the influence of drugs	3.675	1	0.055	2.200	0.983	4.927
Face-sitting	4.968	1	0.026	2.119	1.095	4.103
UNDIFFERENTIATED FAILURE OF CONTROL
SEXUAL ROLE
Active/insertive	4.316	2	0.116	Reference
Versartile	3.928	1	0.047	2.028	1.008	4.081
Passive/receptive	0.244	1	0.622	1.267	0.495	3.238
UNPROTECTED SEXUAL PRACTICES
Face-sitting	8.433	1	0.004	2.696	1.380	5.265

Note: df: degrees of freedom; Sig: significance; OR: odds ratio; 95% CI: 95% confidence interval.

. Specifically, education level, cohabitation situation, sexual role, and unprotected sexual practices have been found to be related to SC, according to the four categories of gender stereotypes.

In the feminine category, those participants who perform face-sitting are 5.7 times more likely to have SC (95%CI: 1.098–30.109; p = 0.038). Those who are positioned in a receptive sexual role present a higher level of interference, showing 8 times more risk (95%CI: 1.374–48.011; p = 0.021), while those who declare a versatile role are 5.3 times more likely to be within this dimension (95%CI: 1.132–24.459; p = 0.034) compared with those who have an insertive sexual role. Cohabitation situation is presented in this case as a protective factor against interference in SC, with those who live alone or in situations other than living with a partner showing a lower risk of interference, although this did not reach statistical significance.

Regarding self-identification with the masculine category of gender stereotypes, those who reach a higher level of education present a 9 times greater probability of developing sexual compulsiveness compared with those who have only completed high school (95% CI: 1.390–58.434; p = 0.021). However, those who have completed professional training appear protected compared with the rest in presenting interference (OR: 0.313; 95% CI: 0.100–0.979; p = 0.046). In the masculine category, those who manifest a receptive sexual role have a 3.6 times greater probability of interference (95% CI: 0.977–13.657; p = 0.054). Although this did not reach statistical significance, bisexual MSM self-identifying with masculine stereotypes have a higher probability of interference (OR: 1.429; 95%CI: 0.050–40.855; p = 0.835), and homosexuals have a lower probability of interference (OR: 0.147; 95%CI: 0.006–40.855, p = 0.253) compared with heterosexual MSM. Finally, those in this category who performed fisting in the last year show a 3.4 times higher risk of presenting interference (95%CI: 1.002–11.771; p = 0.050).

The participants included in the androgynous category were protected from presenting compulsiveness when they lived alone or in a situation other than living with a partner; also, in this case, those who had a homosexual affective–sexual orientation presented a lower probability compared with heterosexuals (OR: 0.061; 95% CI: 0.006–0.645; p = 0.020). Participants involved in chemsex in this category were 2.7 times more likely to have interference (95%CI: 1.009–7.266; p = 0.048).

The participants identifying as undifferentiated who practiced face-sitting had a 3.6 times greater risk of presenting compulsivity overall than those who did not carry out this type of practice (95%CI: 1.435–9.083; p = 0.006). The participants who had sex with drug use presented 2 times more risk (95%CI: 0.983–4.927; p = 0.055) and those who performed face-sitting had 2.1 times greater risk of showing interference (95%CI: 1.095–4.103; p = 0.026). Finally, it is the only category that has shown a relationship with the dimension of failure to control sexual compulsivity in the participating MSM, such that those who reported a versatile role had 2 times more risk (95% CI: 1.008–4.081; p= 0.047), while those who assumed the receptive role were 1.26 times more at risk of presenting control failure (95% CI: 0.495–3.238; p = 0.622) compared with those who had an insertive role. Those who practiced face-sitting were 2.7 times more likely to present control failure (95% CI: 1.380–5.265; p = 0.004).

5. Discussion

The present study aimed to understand the possible relation between sexual compulsivity and the different gender stereotypes with which MSM self-identify, based on information related to sexual and behavioral characteristics. Behavioral characteristics and unprotected sexual practices were found to be related to SC. Bisexual and homosexual participants showed less sexual compulsivity. It seems that considering themselves under an LGBTI label, sheltered by a group of gay and bisexual men (GB), could protect against sexual compulsiveness in these MSM populations, a fact that has already been shown to play a beneficial role in mental health and stress protection in sexual minorities [52]. Although studies related to sexual health have focused more on GB labels than on other MSM [53], these findings indicate that more depth is required in the study of these groups due, in particular, to the existence of more clandestine situations [54] that can lead to greater compulsive behaviors. In line with other studies, the use of information and communications technologies (ICTs) was related to SC [55] and with drug consumption when having sexual practices, and there was also a high association with chemsex [56,57,58]. Surprisingly, we did not find any association between SC and PrEP, unlike previous studies [59,60] and neither was there any association with the existence of previous diagnoses of STD [61].

Regarding unprotected sexual practices, face-sitting and fisting stand out, which, in agreement with other studies, have also been identified as practices especially related to SC and substance use [62,63].

In general, this could be considered the profile of MSM who access this type of study [64]. However, the idea of said profile must be taken with caution if the intention is to extrapolate this information to populations in specific situations of vulnerability that determine their sexual health, their behaviors, and the relationship of all the factors studied with gender stereotypes, spaces in which the greatest stigmatization exists [65,66,67]. Furthermore, the epidemiological term “MSM” is frequently used to refer to sexual behavior alone, regardless of sexual orientation; for example, heterosexual men, who should be included when they sporadically engage in sex with other men, in clandestine spaces, or in rural areas [68].

As others have said, in the study of gender stereotypes in MSM, there are hardly any tools beyond the focus on traditional masculinity [38]. This suggests a barrier to the knowledge of the diverse realities in terms of gender beyond binarism, indicating, as does some previous research, that gender roles are not inherent to sex [69], which should be incorporated into health studies [70].

In order to identify the gender stereotypes in MSM and to analyze the different forms of constructing sexual and affective relations among MSM, BSRI-12 was used in our study, showing excellent psychometric properties for its application in the sample and construction validity, as in other studies [71,72] and in the previous study performed in an MSM population in Spain [38]. The instrument allows us to analyze how self-identified instrumental and/or expressive traits, according to Bem’s proposal [34,36,46], can be related with our study variables. As in previous research, our results demonstrate the association between gender stereotypes and the engagement in sexual risk practices and, specifically in our case, with SC.

In our sample, we found a higher proportion of those who self-identified as undifferentiated (325; 36.9%) and androgynous (283; 32.1%). Related to these, our results are consistent with those of the previously study in MSM gender self-identification in Spain [38] and in other populations [72,73,74]. As 69% of the participants deviate from traditional masculine/feminine self-identification traits, it suggests that in MSM diversity, sexual orientation and gender identity are concepts to be taken together to better understand their sexual health determinants. With regard to gender order [75], it seems to be that when MSM are placed in lower positions of the gender hierarchy of masculinities, engagement in risky practices and SC could be intensively related.

Androgyny is characterized by a combination of both masculine and feminine traits, considered by previous research as being more assertive and capable of adapting one’s own behaviors to different situations using both instrumental and expressive characteristics and, therefore, obtaining better health outcomes [72,76]. Nevertheless, as seems to be the norm for other populations, in MSM, the androgynous category is related to greater sexual compulsivity, as our results suggest. Furthermore, other variables such as cohabitation situation have an impact [77]. Living alone has been found to be related to sexual risk practices [78] by presenting more options of having multiple partners, which have been also related to SC [79].

Even so, according to the model presented, the participants included in the undifferentiated category are the ones most likely to engage in unprotected practices, specifically face-sitting and fisting, in addition to drug-influenced sex. The undifferentiated category describes people whose scores for both masculine and feminine traits are low, that is, those with a lower self-perception of both instrumental and expressive traits. Previous research on MSM also found that undifferentiated participants were more vulnerable to engaging in risk practices [38].

The feminine category showed lower SC in our study, while the masculine category was related to failure of control with progressive age [80]. Therefore, gender stereotypes that are close to hegemonic masculinity have an impact on sexual health and determine the different ways of relating [38].

There are more and more studies that have found a relationship between masculine stereotypes, sexual risk practices, and SC in current contexts about MSM [81,82,83].

Beyond binary gender stereotypes, it is necessary to know how gender demands behave in MSM, since a critical view of the heteronormative sexual standards that continue to persist in the clinical health literature seems to be necessary [84]. The relationships between gender stereotypes, affective–sexual diversity, and health behaviors require a deep insight into the identities, fantasies, behaviors, and specific environments in dissident sex–gender people and in MSM in general [85,86].

Although these analyses have found a significant association between SC and gender stereotype self-identification in MSM, it is worth mentioning that the SCS and the BSRI-12 could dualistically assess the existence of sexual risk behavior in the MSM population.

Researchers and public health service providers should consider reducing the invisibility of sexual education in sexual sex and gender minority groups, addressing sexual compulsivity and gender stereotypes and their impact on sexual behaviors that MSM may engage in. They should also include collectively sexually educating MSM about the potential risks that are associated with sexually compulsive behaviors and promoting a healthier self-identified gender outside the heteronorm and beyond individual responsibility by making practices different from heterosexual ones invisible. Thus, these findings highlight the need for future research that examines how gender stereotypes influence the sexual health of MSM, while also exploring the role that sexual compulsivity may be contributing to this association.

This study has several limitations, such as those derived from the cross-sectional design, which does not allow for the determination of a cause–effect relationship. Also, it must be acknowledged that the sampling was conducted for convenience and the data collected was voluntary and via the internet, which means that many MSM will not have had access to it. Future studies using prospective study designs are necessary to examine the longitudinal nature of the associations found in our research. Although SCS was previously validated in MSM in Spain and the BSRI was also used, this is the first study, to our knowledge, that adds the different components of the applied instruments (SCS; gender, sexual, and behavioral traits) developed in this population. Although this statement can be taken as a weakness, it at the same time shows new ways of understanding the situations related to the health and self-care of MSM in the current changing society in relating ways and opportunities, which represent new challenges for health professionals and policy makers. According to the current state of the art, information obtained from the BSRI-12 allows opportunities to reflect on the dichotomous thought regarding gender by taking into account different gender characteristics in MSM. It also facilitates the identification of vulnerability situations determining sexual risk practices in MSM.

6. Conclusions

Sexual compulsivity is related to gender stereotype self-identification in MSM men in Spain. Sociodemographic, behavioral, and unsafe sexual risk practices are related to SC and also to non-binary gender self-identification in this population. Monthly income is related to SC interference in feminine or expressive traits, while failure of control is related to masculine or instrumental traits. Androgynous is the most frequently self-identified gender among MSM and is related to living alone. The undifferentiated category shows lower scores on both instrumental and expressive gender stereotypes, indicating trends in the most frequent engagement in high-risk practices in MSM.

Sexual compulsivity and gender stereotypes are associated, and a call to action is needed, as well as moving forward heteronormative standards to understand current trends for better sexual health in MSM.